No description available.

In this thesis, a suite of techniques and algorithms is presented to tackle three main tasks. Firstly, many existing image-related approaches (processing or analysis) need to be extended from low-dimensional space (e.g. 2D) to a higher-dimensional space(3D). In addition, they often also need to be improved to achieve better accuracy and more efficiency to enable processing ofmassive volumetric images. Frequently new / techniques or algorithms also need to be developed to cover the gap in these previous requirements. Based on these approaches, the second task is to extract the geometric and topological properties of the pore space directly from 3D images of rock samples. The third task is then to study and to establish the relationship between the microstructure and the macroscopic properties by constructing realistic network structures for network models or by conducting some numerical experiments such as mercury injection etc. In the framework of the methodology presented in this thesis, many commonly used image processing and analysis approaches form the basis of the pore space quantification procedure. These primarily include 3D Euclidean distance transformations, 3D geodesic distance transformations, component labelling (clustering), and morphological operations. Among these techniques, some are either unavailable in 3D discrete space or are of too low-efficiency for handling the huge size of rock samples, and others simply did not exist prior to my work. The next level of the methodology is. to quantify the pore space. In order to process 3D images efficiently thus, firstly, the medial axis (skeleton) of the object (e.g. the pore space) is generated so that simple and compact basic information of the object remains while irrelevant redundant information is neglected in the resultant skeleton image. Having obtained the skeleton of an object, most of the geometric and topological quantities of this object can then be easily derived. After reviewing many existing algorithms, a more accurate and efficient thinning algorithm is presented to meet the specific requirements for the study of pore microstructure. Furthermore, general geometric and topological properties of the pore space are calculated and analysed, including pore size distribution, bond (or node) radii/length/volume, shape factor and coordination number etc. As an important contribution, a novel algorithm to compute the Euler-Poincare characteristic (Euler number) is presented and a new topological descriptor is introduced to overcome the limitations of the Euler number and the coordination number. To validate the methodology and to carry out some basic analysis of the microstructure of porous media, I investigate the geometric and topologic features directly from 3D binary images of rock samples. The volum~tric pore size distribution is obtained, and the frequency of pore inscribed radii (or diameter) is calculated, the shape of cross sections along pore channels is quantified as the shape factor and the corresponding algorithm is created. In this study, many quantities for describing the morphological properties of porous media have been successfully introduced. To carry this novel methodology into the use of network models for the prediction of flow processes, three rock samples are selected and analysed. A new approach is developed for partitioning the pore space into the network of nodes and bonds. This partitioning differs from existing methods and it aims to solve some specific problems which often occur in unconsolidated (high porosity) porous media. Following this some single/multi-phase properties are calculated for these three rock samples, such as absolute permeabilities and relative p.ermeabilities. A number of relations between pore size and the absolute permeability, or between pore connectivity and absolute permeability, are explored. The comprehensive relation between pore size, connectivity and absolute permeabilities is also studied and preliminary results are given. This research has created new tools that will play important roles in the analysis of porous media.

This thesis presents two new sensors for measuring permeability and acoustic properties on slabbed core samples in the petrophysics laboratory. Prototypes were built to demonstrate that the new techniques are viable. The permeability sensor is the first non-contact, high spatial resolution, permeameter ever to have been developed; it has been called a micro-permeameter. It was calibrated using both synthetic and real rock samples. The collected data were compared with results from the industry-standard CoreLab PDPK-400 mini-permeameter. The comparison shows that the new micro-permeameter is much faster than the CoreLab instrument. In addition, the new micro-permeameter is more sensitive to permeability changes and yet still makes more repeatable measurements. The uniqueness of the new micro-permeameter project resulted in the granting of US patent No. 6715341 (see Appendix D). The acoustic sensor developed is laser-based for the excitation stage only, and piezoelectric-based for the detection stage. The laser was tuned to generate acoustic waves in the rock without harming the rock surface. The new experimental method is comparable with exploration seismic surveys, but unique in the petrophysics laboratory. As a result, seismic software may be used to analyse the collected data. A significant discovery was made that not only was the new laser-acoustic system able to identify the standard direct-wave events, but also it reflected and refracted events never identified in petrophysics laboratory before. The standard velocity measurements for the pressure- wave (Yp) and the shear-wave (Ys) may now be made with a far greater accuracy than is possible with the established single-path method. Furthermore, the accuracy may now be measured statistically, rather than as an estimate. Direct measurements of amplitude-versus-offset (AYO) phenomena, which are impossible with the established method, may now be made. The results could be used to calibrate exploration seismic data, with fewer assumptions than the standard approach.

As uncertainty can never be removed from reservoir forecasts, the accurate quantification of uncertainty is the only appropriate method to make reservoir predictions. Bayes’ Theorem defines a framework by which the uncertainty in a reservoir can be ascertained by updating prior definitions of uncertainty with the mismatch between our simulation models and the measured production data. In the simplest version of the Bayesian methodology we assume that a realistic representation our field exists as a particular combination of model parameters from a set of uniform prior ranges. All models are believed to be initially equally likely, but are updated to new values of uncertainty based on the misfit between the historical and production data. Furthermore, most effort in reservoir uncertainty quantification and automated history matching has been applied to non-geological model parameters, preferring to leave the geological aspects of the reservoir static. While such an approach is the easiest to apply, the reality is that the majority of the reservoir uncertainty is sourced from the geological aspects of the reservoir, therefore geological parameters should be included in the prior and those priors should be conditioned to include the full amount of geological knowledge so as to remove combinations that are not possible in nature. This thesis develops methods of geological parameterisation to capture geological features and assess the impact of geologically derived non-uniform prior definitions and the choice of modelling method/interpretation on the quantification of uncertainty. A number of case studies are developed, using synthetic models and a real field data set, that show the inclusion of geological prior data reduces the amount of quantified uncertainty and improves the performance of sampling. The framework allows the inclusion of any data type, to reflect the variety of geological information sources. ii Errors in the interpretation of the geology and/or the choice of an appropriate modelling method have an impact on the quantified uncertainty. In the cases developed in this thesis all models were able to produce good history matches, but the differences in the models lead to differences in the amount of quantified uncertainty. The result is that each quantification would lead to different development decisions and that the a combination of several models may be required when a single modelling approach cannot be defined. The overall conclusion to the work is that geological prior data should be used in uncertainty quantification to reduce the uncertainty in forecasts by preventing bias from non-realistic models.

No description available.

There is a lack of general design methods for water injection schemes. This is reflected in the fact that there exists no handbook for water injection and no commercial reservoir simulators include modules for injectivity damage; even though waterflood simulation is one of their main tasks. This thesis aims at critically evaluating the current approach and suggesting better practises. The objective is to analyse the parameters affecting the performance of matrix water injection wells and critically evaluate their importance in the light of available field history. This includes a review of the injectivity damage mechanisms involved and evaluation of the prediction methods available to quantifY their reliability for real field cases. The following steps are presented: » Review of current understanding of water injection performance. ? » Review of published coreflood experiments with the aim of identifYing both main trends and the key differences. Other potential damage mechanisms are also discussed briefly. » Available injectivity prediction models are evaluated for sensitivities in the input parameters and compared against both core experiments and field performance. » Field information from more than I00 wells, operated by 15 international oil companies was acquired. The information is critically analysed in consistent manner and the main trends identified and compared to the key findings ofthe laboratory and model approaches. The field data proved generally insufficient to provide a firm correlation relating water quality and formation characteristics to injectivity decline. This comparative investigation does however, indicate the uncertainty range in the key parameters involved and does, therefore, result in an improved understanding of the injectivity mechanisms. The key findings are summarised in brief guidelines for best operational practices for water injection. Furthermore, areas of significant inconsistencies, requiring further investigation, are identified and recommendations made as a basis of research activity to fill some of the many gaps in understanding in this important topic.

One of the legacies of mining is the generation of minewater which can be toxic to the environment and requires treating prior to discharge. This can be treated by either passive or active systems, with the latter being most common for high volume, high load minewaters or those that are difficult to treat. Conventional minewater treatment using pH adjustment can precipitate a large volume of voluminous sludge that is expensive to dispose of. This led to the development of the High Density Sludge (HDS) process in the 1960s, which has now become accepted as the best practical method of treating minewaters whilst minimising the volumes of sludge generated. The mechanisms controlling the formation of HDS are not fully understood and as a result there are a number of misconceptions adversely affecting the use of the HDS process. The primary objective of the present research was to explore in greater depth the mechanisms controlling HDS formation and hence dispel some of these commonly held misconceptions. The current understanding of the HDS process was initially established by undertaking a review of information reported in the technical literature. The key operating parameters and concepts were then identified by undertaking a performance review of the 440l/s HDS plant at the former Wheal Jane tin mine in Cornwall, UK. A series of laboratory batch tests and continuous pilot trials were undertaken to assess the importance of these concepts and generate an understanding of the sludge characteristics and properties. The sludge characteristics were investigated by reviewing the sludge settlement (settlement velocity and the ability of the sludge to self compact) and dewatering characteristics. The sludge properties examined were: mineralogy (by X-Ray Diffraction, XRD), morphology (by Scanning Electron Microscopy, SEM, and Transmission Electron Microscopy, TEM) and surface electrical potential (by . measuring the zeta potential). The research has shown that the dominant mechanism for the formation of Type II HDS involves establishing a pH in the Stage I Reactor that gives a negative charge to the recirculated solids. Physical adsorption then ensures heterogeneous nucleation. Control of the pH in the Stage II Reactor ensures removal of the final trace of metals from solution and return of sludge that is capable of achieving the desired pH in the Stage I Reactor. A review of the relevant literature suggests that too high a pH in the Stage I Reactor will favour homogeneous nucleation and hence inhibit HDS formation. The research has also shown that HDS can be formed using non calcium based alkali reagents and that iron is not required for its generation. Synthetic zinc and manganese minewaters produced HDS with the best settling characteristics. The sludge characterisation showed that there is no requirement for the HDS to be crystalline in nature. The presence of species such as calcium and magnesium cations can 'swamp' the process and slow down the formation of HDS, though they do not prevent its production. Finally, the research undertaken has shown that the HDS process enhances (by over forty times) the dewatering characteristics of the sludge generated during minewater treatment, the original purpose for which the process was developed.

This thesis is concerned with the combustion hazard posed when there are accidental releases of methane from plant, particularly within the petrochemical 'exploration and production' industry. At the present time, such explosion hazards are controlled using explosion suppression systems based around Halon 1301 deployment. However, due to its environmental impact, such methods are being reviewed with the objective being their replacement with environmentally friendly alternatives. During the course of this study, the effect of water in the form of vapour and 'fine' mists has been investigated to determine its effectiveness in the containment and control of a potential methane explosion. Laminar flame/water interaction has been studied in considerable detail, and to a lesser extent the interaction with turbulent burning mechanisms has been studied through a demonstration study based on the conclusions of the laminar flame studies the efficiency of water in various concentrations and states (vapour or liquid-droplets) has been appraised. The research studies have also necessitated fundamental studies of droplet formation via supersaturated vapours, within small-scale laboratory test facilities, and methods suitable for generating large-scale sprays deemed suitable for the replacement of Halon systems were then appraised and characterised so as safety systems of the future may be optimised. The laminar test programme illustrated that water is a competitive explosion suppressant capable of extinguishing a fully propagating flame. It was found that water vapour and 'fine' water droplets are most effective in the mitigation of methane-air flames during early flame formation when curvature effects are a predominant factor, with 'fine' water droplets being more effective than vapour at this time. Turbulent experiments demonstrated that water in the form of 'fine' droplets can be used to fully arrest a propagating stoichiometric methane-air explosion, at concentrations probably less than the molar water concentration associated with inerting methane air explosions. A system based on the 'flashing' concept was fully characterised to illustrate that sprays similar to those utilised in the laboratory combustion work can be produced in a full- scale release, leading to the conclusions that Halon based systems may in the near future be replaced by environmentally sound explosion suppression systems that utilise water.

The deposition of paraffin wax solids in pipelines and risers represents a continuing challenge to flow assurance in offshore installations. Wax deposits reduce product throughput, requiring increased energy expenditure to re-establish flow levels. In severe cases, wax deposits can completely block a pipeline. Preventative solutions to the problem such as pipeline insulation, active heating of pipes or chemical dosing with wax inhibitors are not always economically viable, so mechanical removal using a device known as a 'pig' remains an economical solution to the problem of wax removal. A pig is a cylindrical tool that is driven through the pipe by the flow of product, scraping deposits from the pipe wall as it travels. Despite the importance of pipeline pigging to the oil and gas industry, the effectiveness of pigs in removing wax is poorly understood and it is this problem that is addressed by this thesis. One of the first necessities in undertaking this work has been to define the mechanical properties of wax deposits. This has required critical analysis of published material on the subject of wax deposition along with practical experimentation to create representative models of wax deposits that require mechanical removal from pipelines. Previously, studies of wax removal using pigs have assumed the mechanics of the process to be adequately represented by uniaxial compression or simple shear load models. In this work wax removal is analysed using the orthogonal cutting model. This provides a more accurate description of the process as it includes the effect of material after yielding (the chip) on the net wax removal force. Experiments were designed to allow testing of the validity of the orthogonal cutting theory to the pigging process under a variety of conditions. An original contribution from this work is through experimental and theoretical results that are given context through comparison with established metal cutting theory. Through experimentation a specific cutting energy is obtained for wax removal. The results of the wax cutting experiments have identified particular differences between wax cutting and metal cutting regarding the homogeneity of chip formation. These observations have important implications in predicting wax removal forces using mechanical removal tools. Although the affect of removed wax chips on pigging forces has been neglected in theory, it is well known in practice. The fluid used to drive cleaning pigs is often used to produce a jet radiating centrally from the front of the pig intended to blast wax chips away from the pig body, avoiding formation of a 'plug' of wax ahead of the pig. In this study a novel variation of this process in the form of an annular bypass jet is experimentally studied. A semi-empirical model of wax removal using an annular bypass jet has been developed and empirical constants obtained to allow prediction of removal rates for different waxes under various conditions. The new model introduced here allows balancing of pig velocity with wax removal velocity so that a non-contacting wax removal system is obtainable. The bypass-jetting model has been validated using a full-scale trial of the process by industrial sponsors.

The monitoring, interpretation and prediction of gradual ground surface deformation are critical factors in the understanding of earth systems. In many parts of the world, particularly in coastal areas where resources are often easily transportable and where steep cliffs allow access to underlying strata, the patterns of natural ground surface deformation are complicated by mining or extraction activities. To accurately estimate the amount of sea-level rise and Its total affect on, for example, frequency of flooding or salt-water intrusion, the amount of ground surface deformation, either subsidence or uplift, need to be understood in great detail. Ground surface dynamics over an area of contemporary deep mining, IS investigated through two research objectives. A feasibility study of conventional InSAR and Persistent Scatterer InSAR (PSI) in a rural setting was carried out. Rural areas are generally avoided for the application of these techniques for the measurement of gradual ground surface deformation due to the lack of coherence between scenes. The results demonstrate that the new PSI technique significantly outperformed conventional InSAR m the detection of gradual ground surface deformation. However, limitations to the technique were identified in the low density and limited distribution of permanent scatterers. The behaviour of the deformation rate over time appears to be biased to a linear trend. Furthermore, in order to understand the link between underground mining activities and local ground surface response a data-driven model has been developed and evaluated. Based on different mining scenarios, this mode! IS able to estimate the total subsidence in a four dimensional space. It was found that local ground surface deformation can be forecasted accurately, based on an angle of draw and four variables. Five key indicators, which are the extent of die disturbed area, the total period of deformation, the peak rate, the moment of the peak rate and the total deformation, are relevant to understand the impact of underground excavations on the surface and to place it in a wider Earth system.